基于全工况数学模型诊断核电汽轮机热力系统故障的新思路

New Idea for Thermal Parameters Fault Diagnosis of Nuclear Power Steam Turbine System Based on Large-scale Off-design Mathematic Model

利用核电汽轮机热力系统大范围变工况(全工况)数学模型建立核电热力系统故障诊断系统。通过数学模型对对象系统机理的描述,避免了以往神经网络等推理方法中对大量运行数据的需要,更具实用性。由于引入了反映设备运行性能的特性参数概念,通过特性参数随其主导因素的变化关系诊断系统设备故障,避免了依据表面现象判断故障的困难,并使故障诊断结果具有量化标准,能对现场运行和维修提出有益的指导。

Based on the large-scale off-design mathematic model, a thermal parameter fault diagnosis system was set up for nuclear power steam turbine. Due to the description of the mechanism of thermal power system in the mathematic model, the training procedure using large amount of operation data in former diagnosis system was avoided, which made this method more practical. Because the method introduced the concept of Characteristic Parameter and diagnosis the faults by the relationships between the Characteristic Parameter and its Dominant Factor, the difficulty of fault diagnosis based on phenomena was also avoided. Further more the fault diagnosis result is quantitative and could be an ideal guide for the operation and maintenance of the object system.